Synthesis and Characterization of Magnetoelectric Ba7Mn4O15.

We present the synthesis of a novel binary metal oxide material: Ba7Mn4O15. The crystal structure has been investigated by high-resolution powder synchrotron X-ray diffraction in the temperature range of 100-300 K as well as by powder neutron diffraction at 10 and 80 K. This material represents an isostructural barium-substituted analogue of the layered material Sr7Mn4O15 that forms its own structural class. However, we find that Ba7Mn4O15 adopts a distinct magnetic ordering, resulting in a magnetoelectric ground state below 50 K. The likely magnetoelectric coupling mechanisms have been inferred from performing a careful symmetry-adapted refinement against the powder neutron diffraction experiments, as well as by making a comparison with the nonmagnetoelectric ground state of Sr7Mn4O15.

Introduction

Solid-state phases containing Mn–O systems are of great interest due to the variety of functional properties they may exhibit, such as colossal magnetoresistance, ferromagnetism (FM), and ferroelectricity (FE).[1] Multiferroic materials combine the latter two properties in a single material and have gained popularity recently for both their fundamental chemical novelty and their potential applications in devices.[2] If the onset of magnetic and ferroelectric orderings have different mechanisms, the phase is known as a type I multiferroic. These phases typically have different FM and FE ordering temperatures, and the coupling between the two properties is weak. In a type II multiferroic, the magnetic ordering is induced by the ferroelectric ordering or vice versa, leading to a single ordering temperature and strong coupling between the ferroic states. One of the best-known examples of a Mn–O-containing multiferroic is YMn2O5, an orthorhombic Pbam structure comprising edge-shared chains of MnIVO6 octahedra along the c axis that are linked by MnIIIO5 pyramids.[3] The mechanism of the multiferroicity in this phase has been identified as exchange striction, in which the onset of magnetic ordering causes a polar structural distortion.[4,5] All of the resulting components of the electrical polarization lie along the b axis.[6] Another example of a Mn-containing multiferroic is TbMnO3, in which the FE polarization is moderated by the Dzyaloshinkii–Moriya (DI) interaction and is thus quite weak.[7−9] The arrangement of the Mn–O polyhedra in these phases has significant effects on the magnetoelectric coupling, and other multiferroics may be discovered by exploring phases containing unusual Mn–O linkages. A variety of type II multiferroic Mn–O-containing phases may already be found in the literature with ferroelectricity, which is induced by improper mechanisms, magnetic ordering, charge ordering, or orbital ordering.[10−15]

The Sr7Mn4O15 phase was first described by Kriegel et al.; it crystallizes in the P21/c space group and contains face-sharing Mn2O9 octahedral dimers, which share corners to form strings in the ac plane; these strings stack along a to give a quasi-two-dimensional structure.[16,17] The face-sharing dimer motif is uncommon compared to solely corner-sharing systems, appearing in 4H-SrMnO3 and as infinite chains in 2H-BaMnO3, and in perovskites has been shown to appear as a particular result of relative ionic sizes.[18] In Sr7Mn4O15, these dimer units result in strong antiferromagnetic (AFM) interactions between the neighboring Mn4+ sites, a broad peak in the DC magnetic susceptibility and divergence between the field-cooled cooling (FC) and zero-field-cooled (ZFC) warming curves.[19,20] In addition to the broad maximum, we have previously observed divergences between the FC and ZFC susceptibility results for the series Sr7Mn4O15 to Sr3.5Ca3.5Mn4O15 below 175 K, leading us to suggest that the behavior could be explained by weak ferromagnetism (wFM) in which the AFM spins are slightly canted. We propose that the space group in which this is allowed, P21, would also allow local displacements of the oxide anions to produce FE ordering, resulting in a magnetoelectric ground state for the phase.[21] In general, the low symmetry of this structural type means that a large number of ways exist in which the magnetic exchange interactions may break inversion symmetry and hence allow for the magnetoelectric effect. Motived by this fact and previous reports of limited isovalent substitutions on the Sr site, we have prepared the Ba analogue Ba7Mn4O15, which represents a novel binary oxide of Ba and Mn. Our detailed characterization of the magnetic ground state suggests that it may possess magnetoelectric coupling.

Experimental Details

To synthesize Ba7Mn4O15, BaCO3 (99.95%, Alfa Aesar) and MnO2 (99.996%, Alfa Aesar) were ground together in a 7.7:4 ratio (i.e., a 10% molar excess of BaCO3 compared to a stoichiometric reaction) and pressed into a pellet (diameter: 13 mm) under 7.5 metric tonnes of force. The pellet was heated to 900 °C for 48 h, then reground, pressed, and heated to 900 °C again. The grinding, pressing, and heating process was performed five times, and heating was always performed under an atmosphere of flowing N2.

To synthesize Sr7Mn4O15, SrCO3 (99.9%, Sigma-Aldrich) and MnO2 (99.996%, Alfa Aesar) were ground together in a 7:4 ratio and pressed into a pellet (diameter: 13 mm) under 7.5 metric tonnes of force. The pellet was heated to 900 °C for 20 h, then reground, pressed, and heated to 1000 °C for 24 h. The grinding, pressing, and heating process was performed six times under air.

High-resolution powder synchrotron X-ray diffraction experiments were performed at Beamline I11 at Diamond Light Source, with diffraction patterns recorded at 300 and 100 K using the multianalyzer crystal (MAC) for both Ba7Mn4O15 and Sr7Mn4O15, and variable-temperature measurements performed between these temperatures using the Mythen detector for Ba7Mn4O15. The beam wavelength was 0.826831 Å for the Ba7Mn4O15 diffraction experiment and 0.826341 Å for the Sr7Mn4O15, refined using NIST 640c Si standards. Powder neutron diffraction experiments were performed using the GEM instrument at the ISIS Neutron and Muon Source at 80 and 10 K for Ba7Mn4O15 and using the D2B instrument at the Institut Laue-Langevin (ILL) at 300, 100, and 1.5 K for Sr7Mn4O15.[22,23]

Rietveld refinements were performed using TOPAS Academic v6.[24] A starting model for the refinement was based on the crystal structure of the related Sr7Mn4O15 phase with space group 21/. Combined Rietveld refinements were performed using the 100 K dataset from the I11 MAC detector and the 80 K dataset from the GEM experiment. These refinements excluded detector bank 1 and bank 6 from the GEM dataset due to poor signal-to-noise ratios. Refined lattice parameters and atomic coordinates for Ba7Mn4O15 at 300 K are summarized in the Supporting Information (Table S1) isotropic displacement parameters were constrained to be equal for a given atom. Refinements of the low-temperature magnetic structure were performed within the symmetry-adapted formalism of the ISODISTORT suite[25] and as implemented through the linear constraints language of TOPAS. Refined values from these refinements are also summarized in the Supporting Information (Table S2).

Magnetization measurements were performed using a Quantum Design MPMS SQuID magnetometer. Magnetic susceptibility versus temperature data were collected between 10 and 325 K in an applied field of 100 Oe under zero-field-cooled warming and field-cooled cooling conditions. Magnetization versus field curves were collected in applied fields of up to 50 kOe at temperatures between 2 and 300 K.

Results and Discussion

Sr7Mn4O15 has been demonstrated to be flexible to limited substitution by both Ca2+ and Ba2+ cations at the Sr2+ sites.[19,20] Each substituting cation preferentially occupies different Sr sites within the unit cell, with the smaller Ca2+ occupying the Sr1 site and the larger Ba2+ occupying the Sr3 and Sr4 sites (see Figure ). However, substitution levels greater than x = 1 for Sr7–BaMn4O15 have not previously been reported. We found that the use of an inert atmosphere and a 10% molar excess of BaCO3 in the synthesis of Ba7Mn4O15 was necessary for obtaining the desired phase. Any attempt to synthesize the product in air results in oxidation of the Mn4+ cation to Mn5+, forming a product which was identified as the Ba3Mn2O8 phase exclusively.[27] Additionally, performing the reaction at a reduced temperature of 900 °C (compared to literature at 1300 °C) produced the purest material. If an inert atmosphere is used without excess BaCO3 in the reaction mixture, the reaction produces a mixture of Ba7Mn4O15 and a secondary phase which was identified as Ba4Mn3O10.[28] Since Ba7Mn4O15 possesses a greater Ba:Mn ratio than Ba4Mn3O10, we found that a 10% excess of BaCO3 produces a near phase-pure product, with just a small quantity of unreacted poorly crystalline BaCO3 in the diffraction pattern. The inert atmosphere, requisite excess of BaCO3 and exceptionally low synthesis temperature required to stabilize this phase are no doubt contributing factors as to why Ba7Mn4O15 has hitherto remained unidentified.

Figure 1

Unit cell of Ba7Mn4O15 with sites labeled. Purple polyhedra represent Mn2O9 octahedral dimers, green spheres indicate Ba atoms, and red spheres represent O atoms.

Figure shows the result of a Rietveld refinement against the high-resolution powder synchrotron X-ray diffraction data collected at 300 K of Ba7Mn4O15; the unreacted BaCO3 appears as a broad peak around 2θ = 13° with a calculated concentration of ∼12% by weight. This value may be overestimated due to the low crystallinity of the residual reagent. The unit cell of Ba7Mn4O15 is similar to that reported for Sr7Mn4O15, containing the same Mn2O9 dimer units, but with larger lattice parameters resulting from the greater ionic radius of Ba2+ compared to Sr2+. At 300 K, we do not find evidence for the disordering of the Sr(3) and O(6) sites from their high-symmetry positions as described by Vente et al. for Sr7Mn4O15.[20]

Figure 2

Rietveld refinement against powder synchrotron X-ray diffraction (λ = 0.826831 Å) data at 300 K. Blue tick marks indicate reflections for Ba7Mn4O15; green tick marks indicate reflections for BaCO3.

The variations in lattice parameters and the angle β with temperature between 300 and 100 K are shown in Figure . All parameters vary monotonically with temperature, indicating the absence of any structural phase transitions in this temperature range.

Figure 3

Lattice parameters of Ba7Mn4O15 as a function of temperature.

In the case of Sr7Mn4O15, the strings of Mn2O9 dimers were predicted to result in strong magnetic exchange interactions in the bc plane, but weak interactions along a.[20] Previous reports of magnetic susceptibility experiments on Sr7Mn4O15 describe a broad maximum centered around 75–90 K[20,21] with a small upward tail below around 20 K. The FC and ZFC curves diverge from one another at temperatures below this maximum. This behavior has been explained with two different mechanisms: Vente et al. proposed that it resulted from spin glass-like behavior producing clusters of antiferromagnetically ordered spins, which crystallize into true antiferromagnetic order below ∼75 K, whereas we have previously suggested that it might represent a weak FM ordering component, arising from the local symmetry-breaking associated with the aforementioned disorder of the Sr and O sites.[20,21]

We find that the DC susceptibility versus temperature results for Sr7Mn4O15 match well with previous descriptions, with the maximum of the broad feature centered around ∼84 K and the deviation between the ZFC warming and FC curves below this temperature (Figure ). In comparison, the DC magnetic susceptibility results for Ba7Mn4O15 are relatively featureless. We observe a steady upward trend with decreasing temperature between 300 and 50 K. However, on further cooling a clear divergence between the FC and ZFC warming curves is visible, evidencing a possible long-range magnetic ordering transition. The susceptibility obeys the Curie–Weiss law in the 200–300 K range. A fit to this part of the inverse susceptibility in the ZFC warming data yields a μeff = 3.78(7) μB per Mn4+ site for Ba7Mn4O15 that compares favorably against the expected spin-only value of 3.87 μB.

Figure 4

DC susceptibility versus temperature for Sr7Mn4O15 and Ba7Mn4O15. Inset: Curie–Weiss fit between 200 and 300 K for Ba7Mn4O15.

To further investigate the magnetic behavior of Ba7Mn4O15, we performed powder neutron diffraction at 10 and 80 K (i.e., either side of the divergence) using the time-of-flight powder diffractometer GEM, ISIS. Figure shows the result of a combined Rietveld refinement using the powder synchrotron X-ray diffraction data at 100 K and the powder neutron diffraction data at GEM at 80 K (the same temperature points not having been measured). A fit to data from the same bank at 10 K is also included in the Supporting Information (Figure S1). This refinement of the neutron diffraction data at 80 K is well above the suspected magnetic ordering temperature and shows no significant unfit intensity by our nuclear model. Figure shows an enhanced view of the d = 3.1–3.4 Å and d = 4.2–4.9 Å regions of neutron diffraction data from the same bank at 10 K; magnetic Bragg reflections are evident that index as (0 1 2), (1 0 −2), (1 2 −2), and (2 1 −1).

Figure 5

Results of combined Rietveld refinement against GEM and I11 diffraction data at 80 and 100 K, respectively, for Ba7Mn4O15. Visualized is the fit against the data on bank 3 of GEM, though data from all banks was used to generate the model. Blue tick marks indicate reflections for Ba7Mn4O15; green tick marks indicate reflections for BaCO3.

Figure 6

Comparison of the powder neutron diffraction data for Ba7Mn4O15 from GEM bank 3 at 80 and 10 K, showing the development of magnetic Bragg peaks.

To fit these magnetic peaks and solve the magnetic structure, the nuclear structure refined using the combined 80 and 100 K data above was used to produce a .cif file, which was used as a starting model in ISODISTORT for the 10 K neutron data. The nuclear structure was fixed and the only parameters refined in the 10 K models were the components of the magnetic modes. There are two symmetry-inequivalent Mn4+ sites in the asymmetric unit, which share a face in a Mn2O9 dimer. In our refinements, we have constrained the moments of these sites to be antiparallel. This is justified by the strong AFM direct exchange interactions expected for the half-full t2g orbitals. We tested relaxing this constraint in our final model, but this led to neither a significant improvement in the fitting statistics nor a substantial deviation from the imposed antiparallel configuration. We tested models considering only a single magnetic propagation vector k = [0 0 0], as the magnetic intensities can all be indexed on the nuclear cell. At this k-point there are 4 irreducible representations (irreps), transforming as mΓ1+, mΓ2+, mΓ1–, and mΓ2–, according to the notation used with ISODISTORT. Illustrations of the spin configurations of each of these modes can be found in the Supporting Information (Figure S2); mΓ1+ has FM components along [010] and AFM components along [100] and [001], mΓ2+ has FM components along [100] and [001] and AFM components along [010], and both mΓ1– and mΓ2– have only AFM components along [100], [010], and [001]. The calculated components of the magnetic moments along each direction for each of these modes are included in the Supporting Information in the form of “complete modes details” pages found using ISODISTORT.

The results of the models constructed by considering spin orderings that transform as one of these irreps are summarized in Figure . The single-irrep models fail to account for all of the magnetic peaks, fitting either the (0 1 2) and (1 0 −2) reflections or the (1 2 −2) and (2 1 −1) reflections, but not both. We therefore performed refinements in which a binary combination of modes was constrained to be active in either the ac plane or along b, following the symmetry requirements of the unit cell. The three basis vectors spanning each irrep that describe the possible spin orderings correspond with moments aligned along the symmetry-unique direction b or in the ac plane. We tested all possible combinations exhaustively and report our findings in Table . We note that constraining the moments to lie only along c made no difference to the fits, despite the moment being unconstrained within the ac plane by symmetry. The binary combinations of irreps produce both solutions where only the direction of the moment is modulated (that we will refer to as a spin-wave solution) and solutions where only the magnitude of the moment is modulated (spin density wave), which are highlighted in blue and red boxes, respectively, in Table . For a fixed Mn4+ valence state, a spin wave is more physical, so we restrict our discussion to these in what follows. However, the assertions about the magnetoelectric ground state that we present below hold true irrespective of this fact.

Figure 7

Fits to 10 K powder neutron diffraction data from GEM bank 3 at for Ba7Mn4O15 using single-mode models and dual-mode models, showing the failure of individual modes to capture all magnetic Bragg peak intensity. The underfit intensity in the peak at 4.5 Å is due to the BaCO3 impurity, marked with a star symbol.

Table 1

Rwp’s of Refinements of Ba7Mn4O15 Data from GEM Data at 10 K in Which Binary Combinations of Magnetic Modes Were Allowed to Refinea

The boxed region indicates a combination of modes which allow a magnetoelectric (polar) ground state, blue cells indicate spin-wave solutions, and red cells indicate spin density wave solutions.

The combinations of magnetic modes which resulted in the best fit to the data are as follows: mΓ1+ with mΓ1–, both along the c direction, mΓ1+ along c with mΓ2– along b, and mΓ2+ along b with mΓ1– along c. These fits are highlighted in bold in Table . Refinements in which the irreps were also allowed to refine in a were also investigated; we find that constraining the irreps along the c and b directions does not negatively impact the quality of the fit to the data, in line with literature predictions of the magnetic structure of Sr7Mn4O15.[20,21] While the components of the magnetic moments were not constrained to be equal along the b and c lattice directions, they consistently refined to approximately equal values as shown in Figure .

Figure 8

Possible spin configurations for Ba7Mn4O15, where the magnetic moments are constrained to be in the bc plane. The total magnetic moment for the Pc model is 2.46 μB (1.6 μB along b, 1.9 μB along c), and the total magnetic moment for the Pc′ model is 2.34 μB (1.5 μB along b, 1.8 μB along c). The two symmetry-unique Mn sites in P21/c are indicated by blue and red arrows; their coupling is constrained to be AFM as this was found to fit the experimental data best and expected by the strong magnetic direct exchange interactions.

Despite demonstrating fitting statistics that equal the best models, we discard the first of the combinations (mΓ1+ with mΓ1–, both parallel to c) as it results in a spin density wave with unphysical descriptions of the magnitudes of the moments: namely, the calculated magnitude of the moments on half of the Mn sites are equal to 3.5 μB, while the other half of the sites have magnitudes of 0.1 μB. The remaining combinations of modes result in AFM spin-wave structures with magnetic space groups Pc and Pc′. The fits to the data from these models are shown in the right-hand-side column of Figure , and the spin configurations within the unit cell are shown in Figure . We note that the presence of a small BaCO3 impurity results in slight underfitting of the peak at 4.5 Å. These two models differ only slightly in the arrangement of the magnetic moments, and both produce magnetic moment magnitudes of around 2.3–2.4 μB. This is slightly reduced from the maximum expected value of 3 μB, which we attribute to the overlap between t2g orbitals of the Mn centers within the Mn2O9 dimers causing a loss of spin density.

The fact that these two spin configurations are the best-fitting models to our data at low temperature is of significant interest: the joint action of the two magnetic irreps—one of which conserves inversion symmetry (Γ+) and one of which violates inversion symmetry (Γ–)—results in a structural space group in which inversion symmetry is globally broken.[29] Indeed, comparing this type of combination (highlighted in Table with a dark box) with combinations where both modes either conserve or do not conserve inversion symmetry, we find that the combinations resulting in a noncentrosymmetric space group fit the data best in all cases. While we are not able to detect an off-centering of any of the high-symmetry positions in Ba7Mn4O15 in our powder diffraction data, the crystallographic analysis indicates that this phase possesses a ground state in which the noncentrosymmetric space group (Pc) is a direct result of the magnetic ordering, implying that it may be magnetoelectric and/or multiferroic. A table of the space groups which result from the combination of multiple magnetic modes is shown in the Supporting Information (Table S3).

Figure shows magnetization measurements for Ba7Mn4O15 as a function of the applied field. A small but clear hysteresis is evident at temperatures less than 25 K, with a maximum moment of 0.083 μB per Mn4+ at 2 K and 50 kOe. As this is significantly lower than the effective moment in the paramagnetic region, the ordering is likely to be largely AFM in character, with a small FM component. This can be attributed to a slight canting of the magnetic moments along b transforming as the FM irrep mΓ1+, suggesting that there is a small incomplete cancellation of the magnetic moments within the Mn2O9 dimers along this lattice direction. These magnetization results further support the proposed two-irrep magnetic structure and hence also the proposed polar ground state.

Figure 9

Magnetization versus field results for Ba7Mn4O15.

Figure shows one of the sets of allowed polar displacements of the O6 anion. This site and the Ba3 site are the only high-symmetry positions in the 21/ unit cell, and thus we chose one of these to demonstrate a possible multiferroic coupling mechanisms since displacements of these atoms from their average positions will always induce a permanent polarization. The polar displacement transforms as the Γ2– irrep, which is responsible for the displacive distortions that lead to a symmetry reduction from 21/ to Pc (basis = {(1,0,0),(0,1,0),(0,0,1)}, origin shift = (0,1/4,0) with respect to the parent cell). Notably, polar distortions transforming as this irrep appear irrespective of which magnetic space group (Pc or Pc′) is assigned, thus our result is invariant to the ambiguity of the precise magnetic structure with respect to these two space groups. Since the noncentrosymmetric ground state would be induced by the magnetic ordering, Ba7Mn4O15 would be classed as a Type II multiferroic. It is important to emphasize that the displacements indicated in Figure are only one of the possible distortion pathways by which magnetoelectric coupling could be realized. Since our proposed distortions are driven by magnetic ordering, the displacement is expected to be on the order of thousandths of an angstrom. Our experiment is not sensitive to displacements of this magnitude in such a complex structure, and so Figure merely shows the symmetry-allowed character of one such possible displacement, and we can infer nothing more about their magnitude or the relative magnitude of other symmetry-allowed atomic displacements. Additionally, it should be noted that the proposed magnetoelectric ground state arises not only as a result of the canting of the magnetic spins, but is due to the combined action of the two magnetic irreps. While a single magnetic irrep solution can lead to canting in this system, it will never result in a multiferroic ground state.

Figure 10

Symmetry-allowed polar displacements (transforming as Γ2–) of the O6 anion in Ba7Mn4O15 (green polar vectors) in the lower-symmetry Pc space group (basis = {(1,0,0),(0,1,0),(0,0,1)}, origin shift = (0,1/4,0) with respect to the P21/c parent cell).

Finally, we compare the neutron diffraction results for Ba7Mn4O15 with those of Sr7Mn4O15, for which no direct evidence of a magnetoelectric ground state has previously been reported. A powder neutron diffraction pattern for Sr7Mn4O15 is shown in Figure , with the magnetic peaks observed at 1.5 K inset. The most plausible fit to the magnetic peaks is achieved using a model containing only the mΓ2– irrep along the b direction to describe the magnetic moments, consistent with previously published models.[20,21] The spin configuration refined against the data and transforming as the mΓ2– irrep is shown in Figure . This produces a calculated magnetic moment of 2.2 μB consistent with the magnitude observed for Ba7Mn4O15, giving us further confidence in our proposed spin-wave solutions. The two-irrep fits necessary to model Ba7Mn4O15 provide no improvement in the quality of the fit to the magnetic data for Sr7Mn4O15, thus the magnetoelectric ground state we report here is specific to Ba7Mn4O15. We view the Pc (mΓ1+ ⊕ mΓ2–) structure of Ba7Mn4O15 to be the most likely candidate from our refinements, as the mΓ2– mode best describes the magnetic ordering in Sr7Mn4O15, and it is to be expected that the exchange pathways have a high degree of similarity in these two materials.

Figure 11

Rietveld refinement against powder neutron diffraction data collected at D2B for Sr7Mn4O15 at 300 K. Blue ticks indicate reflections for the Sr7Mn4O15 phase; green ticks indicate reflections for a small SrMnO3 impurity. Inset: magnetic reflections observed at 1.5 K (black crosses) and the same region of 2θ at 300 K (green crosses). The magnetic intensity is well modeled by considering only the mΓ2– irrep.

Figure 12

Refined mΓ2– spin configuration for Sr7Mn4O15 (space group P2′1/c), where the magnetic moments are constrained to be in the b direction. The magnetic moment is 2.2 μB. The two symmetry-unique Mn sites in P21/c are indicated by blue and red arrows; their coupling is constrained to be AFM as this was found to fit the experimental data best and expected by the strong magnetic direct exchange interactions.

It is difficult to assign the difference in magnetic behaviors between Ba7Mn4O15 and Sr7Mn4O15 to specific structural features due to the low symmetry of the phases. A selection of the nearest-neighbor distances and angles are summarized in Table S3: while the unit cell obviously expands to accommodate the larger Ba2+ cations, this expansion does not result in exceptional changes to any of the bond lengths or angles between Mn4+ centers. We envisage that substantial future works involving investigating the magnetoelectric ground state of this material, via first-principles calculations, will shed further light on this issue.

Conclusions

We have successfully synthesized the novel ternary compound Ba7Mn4O15. Powder synchrotron X-ray diffraction analysis indicates that this phase remains in the 21/ space group in the 100–300 K temperature range. Powder neutron diffraction and SQuID magnetometry indicate that the phase possesses an AFM ground state below 50 K. Careful analysis of this AFM ground state reveals a small FM component and suggests that a pair of magnetic modes, transforming as distinct irreducible representations, act simultaneously, and in doing so couple to a polar distortion. This is the first experimental evidence that this class of materials can support a multiferroic grounds state, and we hope it will stimulate renewed synthetic effort into preparing structural related materials. Further work on Ba7Mn4O15 should focus on direct characterization of the nature of the polar displacement, confirming and demonstrating the switchability of the polar state and performing synthetic work with the aim of enhancing the magnetoelectric ordering temperature.